SA was originally proposed by Kirkpatrick et al. [46] to design electronic
circuits. This algorithm is based on the analogy of crystal formation
from masses melted at high temperatures and cooled slowly to
allow atoms to align themselves reaching a minimum energy state.
The probability of accepting new solutions is governed by the expression
exp(−ΔE/T), where ΔE is the increment in energy of the new
configuration and T is the temperature. The increment in energy is evaluated
according to the objective function. The initial temperature decreases
geometrically (T = kT) by means of a coefficient of cooling k
once a Markov chain Mc ends. The initial temperature T0 is usually adjusted
followingmethods like that proposed byMedina [47],which consists
of choosing an initial value and checking the percentage of higher
energy solutions accepted. If this percentage is greater than 40%, the initial
temperature is halved; when it is less than 20%, the initial temperature
is doubled. Fig. 4 shows a flowchart of the simulated process. The
algorithm ends when the number of iterations t reaches the maximum
tmax. The procedure can be summarized as follows: